sanjay sarma, steve weis, dan engels mit

52
Low cost RFID and the EPC Sanjay Sarma, Steve Weis, Dan Engels MIT

Upload: others

Post on 23-Apr-2022

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Sanjay Sarma, Steve Weis, Dan Engels MIT

Low cost RFID and the EPC

Sanjay Sarma, Steve Weis,Dan Engels

MIT

Page 2: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• RFID and the Auto-ID Center• Protocols• Security issues

Page 3: Sanjay Sarma, Steve Weis, Dan Engels MIT

Part I Outline

• What and why of RFID• The cost issue• Manufacturing low-cost RFID• Handling the data• Current status

Page 4: Sanjay Sarma, Steve Weis, Dan Engels MIT

RFID System

Page 5: Sanjay Sarma, Steve Weis, Dan Engels MIT

RFID

Page 6: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• What and why RFID• The cost issue• Manufacturing low-cost RFID• Handling the data• Current status

Page 7: Sanjay Sarma, Steve Weis, Dan Engels MIT

why low cost?

4

Page 8: Sanjay Sarma, Steve Weis, Dan Engels MIT

low cost rfid

time

5

10

15

20

die

size

/cos

t, ce

n ts

handling costSilicon: 4c/mm2

Page 9: Sanjay Sarma, Steve Weis, Dan Engels MIT

why is rfid expensive today?

increased chip size greater functionality

reduce functionality(Networking & software)

reduce chip size(handle small chips)

Page 10: Sanjay Sarma, Steve Weis, Dan Engels MIT

Cheap protocol

0102030405060708090

100

905 910 915 920

Frequency (MHz)

Noi

se

Gaussian Fit 0.8912 1002 Meters, 4 KHz,SQW, 1 dBm

Page 11: Sanjay Sarma, Steve Weis, Dan Engels MIT

context-aware router context-aware

router

context-aware router

context-aware router

sensor

context-aware router

context-aware routercontext-aware

routercontext-aware router

context-aware router

context-aware router

context-aware router

sensor sensor sensor

01. 203D2A. 916E8B. 8719BAE03C

Manufacturer 24 bits

Product 24 bits

Serial Number 40 bitsHeader 8 bits

the hypothesis or bet

• Place unique number on tagElectronic Product Code, EPC64 bit, 96 bit, and upwards

• Develop manufacturing technology for small chips and tags

• Move data on the networkNetwork service for resolving EPCNetwork architecture for gathering and routing data

Page 12: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• What and why RFID• The cost issue• Manufacturing low-cost RFID• Handling the data• Current status

Page 13: Sanjay Sarma, Steve Weis, Dan Engels MIT

Low cost RFID

AntennaManufacture

Antenna/ICAssembly

Conversionto Package

Endusers

ICManufacture

20¢ 5¢5¢ 20¢

ICDesign

$X MillionNumber of tags

1-2¢ 1¢1¢ 1¢

ICDesign

$X MillionBillions

Page 14: Sanjay Sarma, Steve Weis, Dan Engels MIT

Slicing and Dicing

• Standard saw-dicing wasteful• Instead, use separation

by thinning

C. Landesberger, S. Scherbaum, G. Schwinn, H. Spöhrle: “New Process Scheme for Wafer Thinning and Stress-free Separation of Ultra Thin IC’s,” Proceedings of Microsystems Technologies 2001, Mesago, Stuttgart, pp. 431-436, 2001.

Page 15: Sanjay Sarma, Steve Weis, Dan Engels MIT

Antenna

• Screen printing• Etching• Forming

M. Feil, C. Adler, G. Klink & M. König: “Interconnection Techniques for Ultra Thin ICs and MEMS Elements”, Proceedings of Microsystems Technologies 2001, Mesago, Stuttgart, pp. 437-442, 2001.

Page 16: Sanjay Sarma, Steve Weis, Dan Engels MIT

vibratory Assembly

Chip Assembly

Chip Design

Label Converting

Inlet Assembly

Silicon Manufacturing

Antenna Manufacturing

End UserLabel / Tag Manufacturing

Wafer Treatment

Orientation Check

Courtesy Philips

Page 17: Sanjay Sarma, Steve Weis, Dan Engels MIT

vibratory assembly

Chip Assembly

Chip Design

Label Converting

Inlet Assembly

Silicon Manufacturing

Antenna Manufacturing

End UserLabel / Tag Manufacturing

Wafer Treatment

Vacuum cylinder

Courtesy Philips

Page 18: Sanjay Sarma, Steve Weis, Dan Engels MIT

conversion

• Paper/package/label industry expertise• Scales well with mass production• Capital equipment expenditure

software

paperhardware

Page 19: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• What and why RFID• The cost issue• Manufacturing low-cost RFID• Handling the data• Current status

Page 20: Sanjay Sarma, Steve Weis, Dan Engels MIT

Architecture: Local

Reader

01.203D2A.916E8B.8719BAE03C

Tag Local database

Reader

Local network

DataProcessing

Page 21: Sanjay Sarma, Steve Weis, Dan Engels MIT

Architecture: Global

01.203D2A.916E8B.8719BAE03C

Reader

Local database

Reader

Local network

Local system

ONS

Internet <PML><TIME=2000.4.28:10:05.05HRS>

<EPC= 01.203D2A.916E8B.8719BAE03C ><TEMPERATURE=15 DEG C>

</EPC></TIME>

</PML>

01.203D2A.916E8B.8719BAE03C

????

Quality control specialist

XQL

PML Server

18.72.100.100

18.72.100.100

PML

<PML>……...<MaximumTempearture>

40 DEG C</MaximumTempearture><MinimumTemperature>

5 DEG C</MinimumTempearture>

</PML>

DataProcessing

Page 22: Sanjay Sarma, Steve Weis, Dan Engels MIT

23AB.36C2.AB21.6733

Tag

Reader

Local database

Reader

Local network

Local system

ONS

Internet<PML>

<TIME=2000.4.28:10:05.05HRS><EPC= 01.203D2A.916E8B.8719BAE03C >

<TEMPERATURE=15 DEG C></EPC>

</TIME></PML>

01.203D2A.916E8B.8719BAE03C

????

Quality control specialist

XQL

PML Server

18.72.100.100

18.72.100.100

PML

<PML>……...<MaximumTempearture>

40 DEG C</MaximumTempearture><MinimumTemperature>

5 DEG C</MinimumTempearture>

</PML>

DataProcessing

01.203D2A.916E8B.8719BAE03C

Inference

Temperature OK

Page 23: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• What and why RFID• The cost issue• Manufacturing low-cost RFID• Handling the data

• Current status

Page 24: Sanjay Sarma, Steve Weis, Dan Engels MIT

Status of center

ResearchRFID/routing software technology: MIT & AdelaideManufacturing /Control Applications: Cambridge

StandardsAir-interface between reader and tagsSoftware for handling/routing data

Sponsorship48 sponsors4 continents

Page 25: Sanjay Sarma, Steve Weis, Dan Engels MIT

field trial

COCA COLA BOTTLERCLEVELAND, TN

JOHNSON & JOHNSONOLIVE BRANCH, TN

KRAFT FOODSFORTH WORTH, TX

GILLETTE DIST. CENTERCHICAGO IL

PILOT TEST FACILITYBENTONVILLE, AR

WAL-MART DEPOTBENTONVILLE, AR.

WAL MART STORECLEVELAND, TN

Warehouse Retail Floor Staging Area Retail Floor

SAM'S DEPOTKANSAS CITY, MO

WAL-MART STOREBROKEN ARROW, OK

SAM'S STORETULSA

P & G FACTORYCAPE GIRADEAU , MO

UNILIVER DIST. CENTERBALTIMORE, MD

P & G DCIOWA CITY, IO

Pilot facility is being used as a mini warehouse

Page 26: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• RFID and the Auto-ID Center• Protocols• Security issues

Page 27: Sanjay Sarma, Steve Weis, Dan Engels MIT

components

• Signaling• Anti-collision• Functions

Page 28: Sanjay Sarma, Steve Weis, Dan Engels MIT

Things to keep in mind

• You will not read one tag: you will read many!

• Bandwidth becomes an issue

Page 29: Sanjay Sarma, Steve Weis, Dan Engels MIT

Line codes

Page 30: Sanjay Sarma, Steve Weis, Dan Engels MIT

trade-offs

Probability of error

Bandwidth

Collision detection

Page 31: Sanjay Sarma, Steve Weis, Dan Engels MIT

modulation

• Amplitude Shift Keying• Phase shift keying• Frequency shift keying

100%

10%

Low Time

45%

45%

Reader to Tag ModulationUHFEnvelope

Page 32: Sanjay Sarma, Steve Weis, Dan Engels MIT

Bandwidth

cos(ωc)cos(ωs)

cos(ωc+ ωs) + cos(ωc-ωs) 2

=

ωcωc- ωs ωc+ ωs

Page 33: Sanjay Sarma, Steve Weis, Dan Engels MIT

bandwidth

Page 34: Sanjay Sarma, Steve Weis, Dan Engels MIT

components

• Signaling• Anti-collision• Functions

Page 35: Sanjay Sarma, Steve Weis, Dan Engels MIT

anticollision

• Multiple tags in the field• Need to be sorted• Tags relatively dumb

Page 36: Sanjay Sarma, Steve Weis, Dan Engels MIT

aloha protocols

State!

Page 37: Sanjay Sarma, Steve Weis, Dan Engels MIT

tree walking

0

00 01

000 001 010 011

1

10 11

100 101 110 111

000 001 010 011 100 101 110 111

Page 38: Sanjay Sarma, Steve Weis, Dan Engels MIT

Aloha vs tree-walking

Reader Tag

Symmetric

Reader Tag

Asymmetric

Aloha schemes13.56 MHz ISM

Tree schemes915 MHz ISM

Page 39: Sanjay Sarma, Steve Weis, Dan Engels MIT

anticollision

• The backbone function • Extract unique ID• Must leverage for any security scheme

Page 40: Sanjay Sarma, Steve Weis, Dan Engels MIT

components

• Signaling• Anti-collision• Functions

Page 41: Sanjay Sarma, Steve Weis, Dan Engels MIT

Functions

• Write address• Lock address• Pointer• Read ID (anti-collision)• Read payload• Write payload• Sleep• Wake• Destroy

One time

Page 42: Sanjay Sarma, Steve Weis, Dan Engels MIT

How to interpret standards

Page 43: Sanjay Sarma, Steve Weis, Dan Engels MIT

outline

• RFID and the Auto-ID Center• A peek at the protocol• Security issues

• Discussions with: Dan Engels, Peter Cole, Steve Weiss, Ron Rivest

Page 44: Sanjay Sarma, Steve Weis, Dan Engels MIT

Does protocol compromise privacy?

Not necessarily. Your choice.

You can destroy the tag and opt out

or

You can keep tag for later use(physics is your friend)

Page 45: Sanjay Sarma, Steve Weis, Dan Engels MIT

mass hijack of tags

• Could happen in destroy or re-programming

• Physics our friendBandwidth limited: 200 tags a second anti-collisionDestroy must be individually addressedSo it takes time to killSurveillance

Page 46: Sanjay Sarma, Steve Weis, Dan Engels MIT

For the future: issues

• Tags are light-weight

• Anyone can read the tags (promiscuity)• The same number shows up all the time• Channel is open and shared

01. 203D2A. 916E8B. 8719BAE03C

Manufacturer 24 bits

Product 24 bits

Serial Number 40 bitsHeader 8 bits

Page 47: Sanjay Sarma, Steve Weis, Dan Engels MIT

problem: unique and promiscuous

Kill Serial number?• Product still readable• Person can be tracked by constellation

Personalize the number?• Repeated reads yield same number• You could still be tracked by constellation

Page 48: Sanjay Sarma, Steve Weis, Dan Engels MIT

using keys – iteration 1

• Tag, reader share a key• They challenge each other• They establish trust• The communicate

But wait:There still needs to be a unique number for anticollisionKey management problem

Page 49: Sanjay Sarma, Steve Weis, Dan Engels MIT

rolling number

Temporary_number = Key{ID|nonce}Perform anticollision on Temporary_numberExtract ID

Advantages:Promiscuous, but who caresCan’t trackNo privacy issue

Disadvantage:Still a key management problem

Acknowledgement: Peter Cole, Ron Rivest

Page 50: Sanjay Sarma, Steve Weis, Dan Engels MIT

keyless approach

lock=hash(key)

Tag knows hash and lockLock used as ID and for anticollision

Administrative functionsReader provides keyTag computes computed_lockIf computed_lock == lock,

then tag unlocked, ready for administration until new lock

Page 51: Sanjay Sarma, Steve Weis, Dan Engels MIT

administering the tag

ProblemsAir interface always vulnerableGrey area of ownership in retail

SolutionsPhysical contact for reprogramming Physical contact reset of memory

(Resurrected duckling, University of Cambridge)

Page 52: Sanjay Sarma, Steve Weis, Dan Engels MIT

conclusions

• RFID is here• As more functionality goes on RFID, security

challenges• 0.2 milli-cents per gate, cost is paramount• Minimize data on tag